M3R-Bench - Target Prediction: leaderboard
M3R-Bench (Jiang et al., 2026) contains 1,000 human-annotated image-text pairs from social media and advertisements for multimodal metaphor understanding, where each model jointly predicts metaphor presence, target and source concepts and sentiment. This board scores the predicted target concept against the annotation by BERTScore F1 on a 0-100 scale; higher is better.
Metric: BERTScore F1 (0-100) of the predicted central target concept against the human annotation, predicted jointly with metaphor presence, source and sentiment. Source: arxiv.org. Saturation forecast: Around July 2028. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 72.39 | #741 |
| 2 | Claude Sonnet 4.6 | 72.09 | #158 |
| 3 | GPT-5.5 | 68.74 | #48 |
| 4 | Qwen 3 VL 8B Instruct | 67.89 | #982 |
| 5 | Qwen 3 VL 32B Instruct | 67.54 | #582 |
| 6 | Kimi K2.5 | 63.17 | #252 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=m3r-bench-target-prediction · How It Works · Data refreshed daily, snapshot 2026-11-12.